[hpsdr] Alex Filter Switching Frequencies

[Phil Harman]

All,

Good to see that Alex is making its way into users hands and appears to be 
working OK.

We need to review just how the various filters on the Alex boards are 
selected .

When I first wrote the Mercury FPGA code that drives Alex,  PowerSDR was 
using a 9kHz IF.  Hence the frequency sent to the HPSDR boards was the 
indicated frequency less 9kHz.  In the Mercury FPGA code I added 9kHz to the 
frequency being sent by PowerSDR and used this to select the approproate HPF 
and LPF pair.

The idea was that the 'intelligence' as to what filters to select was in the 
FPGA code and there was no need for the PC code to send anything more than 
the required frequency - which it did by default.

I gather that the latest versions of PowerSDR use a 0Hz IF in which case the 
filter selection will be incorrect at the high end of some bands. A while 
ago I did try setting the IF frequency to 0Hz in PowerSDR but from memory I 
found the AM demodulator was distorting at zero beat. I can't reproduce this 
problem so assume it has been fixed.

We have a number of options in relation to how the filters are selected - 
these include:

1. Simply remove the 9kHz offset code in the Mercury FPGA code and switch 
filters at the current frequencies. The current logic is:-

For the HPFs
if (frequency < 1.5MHz)  bypass

else if (frequency < 6.5MHz)  1.5MHz HPF

else if (frequency < 9.5MHz) 6.5MHz HPF

else if (frequency < 13MHz)  9.5MHz HPF

else if (frequency < 20MHz)  13MHz HPF

else 20MHz HPF

and for the LPFs

if (frequency > 29.7MHz) 6m LPF

else if (frequency > 21.45MHz) 12/10m LPF

else if (frequency > 14.35MHz) 17/15m LPF

else if (frequency > 7.3MHz) 30/20m LPF

else if (frequency > 4MHz) 60/40m LPF

else if (frequency > 2MHz)  80m LPF

else 160m LPF

2. Remove the 9kHz offset and alter the frequencies that the filters switch 
at. This would help MARS operators.

3. Remove the filter slection from the FPGA code and instead allow the PC 
code to specifically select the filters. This could be simply linked to the 
Band selected or a full GUI similar to that used to select the Penny(Lane) 
open collector outputs or the Alex antenna selections.

When working through the above we also need to consider split/full duplex 
operation and the use of transverters.

Perhaps we could have a preliminary discussion following the Teamspeak 
session on Friday (USA) evening followed by some brainstorming via the 
reflector?

73 Phil....VK6APH







 


 1315570198.0